Hi John Walliker; Re Denali Databahn. Personally, I've never used it, never heard of it. But I haven't been done a chip in a while, just FPGAs. Ask Scumbria.
Stuff like that is necessary for Rambus to get design wins, but it is not sufficient. When it all comes down to the rubber meeting the road, Rambus is simply not getting enough design wins to become the next mainstream memory. DDR is. RDRAM was. There are still RDRAM design wins from 1999 squirting out into production, but the new design wins are DDR.
Since you're stuck off in software land, you don't have a front row seat to watch this happen, but take a look at the DDR and RDRAM chipsets for the PC compatible in this (probably the most complete on the web) list:
users.erols.com
If RDRAM were still alive, someone other than Intel (who had a contractual obligation to use no other) would have chipsets on the drawing board for RDRAM. Look through the above list, there are zero. The only non Intel chipset that uses RDRAM was from ALi, and has since been canned.
-- Carl
P.S. Pardon me if I ignore this topic (i.e. RDRAM vs DDR) for a while. Wait until after the lawsuit stuff is over, which is far more exciting, and I'll talk with you about it. By the way, did you get around to calling up Infineon and order a copy of their datasheet for the x32 300MHz DDR chip? The one that in a single package provides 50% more bandwidth than a Rambus channel? Quite a part, I'd say, an a pretty stong indication that DDR isn't suffering from much of what used to be the vaunted "granularity" problem. |
